[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16523748#comment-16523748
]
Hudson commented on HBASE-20403:
Results for branch master
[build #377 on
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16523216#comment-16523216
]
Hudson commented on HBASE-20403:
Results for branch branch-2.0
[build #474 on
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16523193#comment-16523193
]
Hudson commented on HBASE-20403:
Results for branch branch-2
[build #908 on
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16522743#comment-16522743
]
Andrew Purtell commented on HBASE-20403:
I don't believe branch-1 is affected because of locking
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16522725#comment-16522725
]
Todd Lipcon commented on HBASE-20403:
-
I would guess it's not affected because it has locking in the
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16522721#comment-16522721
]
Mike Drob commented on HBASE-20403:
---
is branch-1 affected by this also? I'd imagine yes since I don't
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16522717#comment-16522717
]
Andrew Purtell commented on HBASE-20403:
Thanks for the fix and the commit [~tlipcon]
>
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16522693#comment-16522693
]
Mike Drob commented on HBASE-20403:
---
bq. Got a report back from an internal test cluster who was
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16522682#comment-16522682
]
Todd Lipcon commented on HBASE-20403:
-
Got a report back from an internal test cluster who was
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16520778#comment-16520778
]
Ted Yu commented on HBASE-20403:
+1
> Prefetch sometimes doesn't work with encrypted file system
>
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16520694#comment-16520694
]
Todd Lipcon commented on HBASE-20403:
-
Sure, I filed HADOOP-15557
> Prefetch sometimes doesn't work
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16520629#comment-16520629
]
Mike Drob commented on HBASE-20403:
---
bq. I think on the HDFS side (Hadoop common, really), we should
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16519982#comment-16519982
]
Hadoop QA commented on HBASE-20403:
---
| (/) *{color:green}+1 overall{color}* |
\\
\\
|| Vote ||
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16519874#comment-16519874
]
Todd Lipcon commented on HBASE-20403:
-
OK. New revision fixes the checkstyle. If someone out there
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16519750#comment-16519750
]
Andrew Purtell commented on HBASE-20403:
It should be ok for prefetch to be slower. It is meant
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16518879#comment-16518879
]
Todd Lipcon commented on HBASE-20403:
-
bq. even though the underlying, wrapped DFSInputStream seems
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16518787#comment-16518787
]
Wei-Chiu Chuang commented on HBASE-20403:
-
Excellent Todd!
Also note that the Hadoop
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16518768#comment-16518768
]
Hadoop QA commented on HBASE-20403:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16518684#comment-16518684
]
Todd Lipcon commented on HBASE-20403:
-
I'm guessing maybe this is partially due to HBASE-17917. It
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16518682#comment-16518682
]
Umesh Agashe commented on HBASE-20403:
--
+1 for the patch! Nice. Thanks [~tlipcon]!
> Prefetch
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16518681#comment-16518681
]
stack commented on HBASE-20403:
---
Thanks [~tlipcon]. Nice. Makes sense (especially given you can repro).
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16517805#comment-16517805
]
Todd Lipcon commented on HBASE-20403:
-
Hello from the peanut gallery!
Looking at the implementation
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16510447#comment-16510447
]
Wei-Chiu Chuang commented on HBASE-20403:
-
This bug seems to have different facets:
{noformat}
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16443242#comment-16443242
]
stack commented on HBASE-20403:
---
Hmm. That makes sense [~jojochuang] and [~apurtell]. In our cache we have
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16439961#comment-16439961
]
Andrew Purtell commented on HBASE-20403:
AFAIK, prefetch does a pass over all blocks of the
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16439957#comment-16439957
]
Andrew Purtell commented on HBASE-20403:
What about prefetch triggers this specifically?
>
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16439949#comment-16439949
]
Umesh Agashe commented on HBASE-20403:
--
Out of 4 ITBLL runs, this stack showed up in log files for 3
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16438210#comment-16438210
]
Andrew Purtell commented on HBASE-20403:
I came to also suggest it's an HDFS bug, and edge case,
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16436397#comment-16436397
]
Umesh Agashe commented on HBASE-20403:
--
Its possible. AFAIK, there is an optimization in the HBase
[
https://issues.apache.org/jira/browse/HBASE-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16436334#comment-16436334
]
Wei-Chiu Chuang commented on HBASE-20403:
-
This looks like a HDFS bug more than a HBASE one.
30 matches
Mail list logo